Re: [HACKERS] statement level trigger causes pltcl, plpython

Tom Lane wrote:
> BTW, one other stylistic nit: I don't think comments like
>
> /* internal error */
> elog(ERROR, "unrecognized OP tg_event: %u", tdata->tg_event);
>
> are really necessary. In the brave new ereport world, any elog(ERROR)
> call is an internal error by definition --- if it isn't, you should be
> using ereport. So the use of elog is sufficient documentation. IMHO
> anyway.

Yeah, I agree. During the conversion, I did that mostly to show that it
was a conscious decision. Going forward, every new elog (versus ereport)
*is* a conscious decision, or at least should be.

Joe